June 2026 platform release notes
These features and SAP Databricks platform improvements were released in June 2026. The cloud the release applies to is indicated for each release note.
Releases are staged. Your SAP Databricks account might not be updated until a week or more after the initial release date.
Anthropic Claude Sonnet 5 now available as a Databricks-hosted model
June 30, 2026 | Applies to: 
AWS GCP Azure
Model Serving now supports Anthropic Claude Sonnet 5 as a Databricks-hosted model. Designed for coding, agentic workflows, and professional work at scale, Claude Sonnet 5 delivers near-Opus-level intelligence with Sonnet's cost efficiency and speed. You can access this model using Foundation Model APIs.
Customers are responsible for ensuring their compliance with applicable terms.
Parquet v2 for Delta Lake tables is generally available
June 29, 2026 | Applies to: AWS GCP Azure
Parquet v2 is now generally available for Delta Lake tables. Parquet v2 uses advanced encodings, v2 data page headers, and INT64 timestamps to improve query performance and reduce table sizes. To use, set the delta.parquet.format.version table property to 2.12.0 or enable automatic upgrades.
Metric view parameters are now in Preview
June 26, 2026 | Applies to: AWS GCP Azure
You can now define parameters in metric views to bind values at query time, serving many query variants from a single definition instead of maintaining a separate metric view per variant.
Data Classification is now available by default for some workspaces with the compliance security profile enabled
June 23, 2026 | Applies to: AWS GCP Azure
Data Classification is now available by default for workspaces with the compliance security profile enabled, and either HIPAA controls or no specific compliance standard selected.
Block specific internet destinations in network policies (Preview)
June 23, 2026 | Applies to: AWS GCP Azure
You can now block internet destinations from serverless workloads through the network policies REST API. Blocked destinations are enforced regardless of the policy's network access mode.
The Discover page is now in Preview
June 22, 2026 | Applies to: AWS GCP Azure
The Discover page is now in Preview. Discover is a curated browsing experience that helps users find data assets governed by Unity Catalog without requiring deep knowledge of catalog hierarchies. Use domains and subdomains to organize tables and dashboards by business area, highlight key assets with custom curation, and explore AI-recommended assets.
Enable access requests across your Unity Catalog metastore (Preview)
June 22, 2026 | Applies to: AWS GCP Azure
You can now enable access requests at the Unity Catalog metastore level. This feature also lets you set a metastore-level destination that's inherited by all catalogs, schemas, and other objects without their own destination. By enabling metastore-level access requests, users can request access to objects even when no destination is explicitly configured.
Legacy compute_preview system schema deprecated
June 22, 2026 | Applies to: AWS GCP Azure
The legacy compute_preview system schema is deprecated and will be removed from the system catalog. Use the system.compute schema instead.
Publishing a Unity Catalog catalog to Microsoft Fabric is in Preview
June 18, 2026 | Applies to: Azure
You can publish an entire Unity Catalog catalog from SAP Databricks to Microsoft Fabric as a read-only mirrored catalog. Fabric users can query the mirrored Unity Catalog tables directly.
Improved workspace files read performance
June 16, 2026 | Applies to: AWS GCP Azure
Common file access patterns are now detected for workspace files, and directory metadata and file contents are prefetched in the background. Workloads that read many files, such as find, grep, Python imports, and Git operations, run much faster.
ai_query is now generally available
June 15, 2026 | Applies to: AWS GCP Azure
The ai_query function is now generally available. This general-purpose AI Function lets you query any supported AI model directly from SQL or Python.
Configure the recovery period for dropped managed tables (Preview)
June 15, 2026 | Applies to: AWS GCP Azure
You can now configure the recovery period for dropped Unity Catalog managed tables. By default, tables are recoverable for 7 days after being dropped using the UNDROP TABLE command. You can set the recovery period to 0 hours (to immediately purge dropped data) or between 7–30 days at the catalog or schema level. A longer period provides added protection against accidental drops of critical data, while a shorter period reduces cloud storage costs for workloads that frequently create and drop tables.
Auto time-to-live for automatic row deletion is generally available
June 12, 2026 | Applies to: AWS GCP Azure
Auto time-to-live (auto-TTL) is now generally available. Auto-TTL automatically deletes rows from Unity Catalog managed Delta Lake tables and managed Apache Iceberg tables based on a DATE or TIMESTAMP column.
Converting a partitioned table to liquid clustering is generally available
June 12, 2026 | Applies to: AWS GCP Azure
You can now convert an existing partitioned Delta Lake table to liquid clustering with ALTER TABLE ... REPLACE PARTITIONED BY WITH CLUSTER BY. Conversion minimizes reader and writer downtime and supports both external and managed tables.
Faster inference tables for custom model serving endpoints
June 12, 2026 | Applies to: AWS GCP Azure
Inference tables for custom model serving endpoints now ingest through endpoint telemetry. Request and response payloads are delivered to a Unity Catalog inference table view within seconds, with configurable sampling. Existing inference tables are not affected. Enable inference tables and telemetry together in the AI Gateway section of the endpoint UI.
Endpoint telemetry for custom model serving endpoints is generally available
June 11, 2026 | Applies to: AWS GCP Azure
Endpoint telemetry for custom model serving endpoints is now generally available. Configure endpoint telemetry to persist OpenTelemetry logs, traces, and metrics from your endpoints to Unity Catalog tables.
ai_extract is now generally available
June 11, 2026 | Applies to: AWS GCP Azure
The ai_extract function is now generally available. This function extracts structured data from text and documents according to a schema you provide, supporting nested objects, arrays, type validation, citations, and confidence scores.
ai_classify is now generally available
June 11, 2026 | Applies to: AWS GCP Azure
The ai_classify function is now generally available. This function classifies text content according to custom labels you provide, with support for label descriptions, global instructions, and multi-label classification.
Delta Sharing is now OpenSharing
June 10, 2026 | Applies to: AWS GCP Azure
Following the release of open-source OpenSharing, Delta Sharing is now OpenSharing. OpenSharing is the new open standard for sharing data and AI assets across platforms and organizations, available both on GitHub and natively on Databricks.
Automatic identity management for Okta (Preview)
June 9, 2026 | Applies to: AWS GCP
Automatic identity management now supports Okta in Preview. You can sync users and groups from Okta into SAP Databricks without configuring SCIM provisioning. When enabled, you can search for Okta users and groups in identity federated workspaces and add them to your workspace. SAP Databricks uses Okta as the source of record, so changes to group memberships in Okta are respected in SAP Databricks.
Anthropic Claude Fable 5 now available as a Databricks-hosted model
June 9, 2026 | Applies to: AWS GCP Azure
Databricks Model Serving now supports Anthropic Claude Fable 5 as a Databricks-hosted model. Claude Fable 5 is built for autonomous knowledge work and coding, designed to handle long-running, complex, and asynchronous tasks with less need for human intervention.
To access this model, use Foundation Model APIs pay-per-token.
Identity federation enabled by default for all new workspaces
June 9, 2026 | Applies to: AWS GCP Azure
Over the next couple of weeks, identity federation is being enabled by default for all new SAP Databricks workspaces, including workspaces that are not attached to a Unity Catalog metastore. This change does not affect existing workspaces.
Inbound Private Link for performance-intensive services is now generally available
June 8, 2026 | Applies to: AWS GCP Azure
Inbound Private Link for performance-intensive services is now generally available.
External lineage is now generally available
June 8, 2026 | Applies to: AWS GCP Azure
External lineage (formerly Bring Your Own Lineage) in Unity Catalog is now generally available. You can register assets that live outside SAP Databricks, such as Salesforce or MySQL sources and Tableau or Power BI dashboards, so that a single lineage graph spans your full data flow.
Foundation model Unity Catalog permissions is generally available
June 5, 2026 | Applies to: AWS GCP Azure
Foundation model Unity Catalog permissions is now generally available. This feature requires enablement by your Databricks account team.
Account admins can use Unity Catalog permissions on the system.ai schema to control which Databricks-hosted foundation models your organization can access. Permissions are enforced across pay-per-token, provisioned throughput, and batch inference (AI Functions) workloads.
Vector Search is now AI Search
June 1, 2026 | Applies to: AWS GCP Azure
Vector Search has been renamed to AI Search. You can now create full text search indexes without any vectors or embeddings needed.
Column popularity is now available in the Catalog Explorer
June 1, 2026 | Applies to: AWS GCP Azure
Column popularity is now available in the Catalog Explorer UI. To find the most popular columns in a table, go to the table's Overview tab. Popularity is determined by the number of queries against the table that read from a given column.